The authors probe the accuracy of linear ridge regression employing a three-body local density representation derived from the atomic cluster expansion. The authors benchmark the accuracy of this framework in the prediction of formation energies and atomic forces in molecules and solids. They find that such a simple regression framework performs on par with state-of-the-art machine learning methods which are, in most cases, more complex and more computationally demanding. Subsequently, the authors look for ways to sparsify the descriptor and further improve the computational efficiency of the method. To this aim, the authors use both principal component analysis and least absolute shrinkage operator regression for energy fitting on six single-element datasets. Both methods highlight the possibility of constructing a descriptor that is four times smaller than the original with a similar or even improved accuracy. Furthermore, the authors find that the reduced descriptors share a sizable fraction of their features across the six independent datasets, hinting at the possibility of designing material-agnostic, optimally compressed, and accurate descriptors.
Claudio Zeni1, Kevin Rossi, Aldo Glielmo and Stefano de Gironcoli J. Chem. Phys. 154, 224112 (2021)
Type of paper:
© 2022, The Author(s)